Extravascular optical coherence tomography of cerebral vessel walls in vivo

Purpose Evaluation of extravascular, microscope integrated OCT (iOCT) as an in vivo imaging modality of cerebral blood vessels and as an intraoperative imaging method. Methods Microscope integrated optical coherence tomography of major cerebral arteries (n = 13) and superficial sylvian veins (n = 5) and one incidental cerebral vasospasm (n = 1) in (n = 10) patients. Post procedural analysis of OCT volume scans, microscopic images and videos during the time of scan as well as measurements of the diameter of vessel walls and its layers with an accuracy of 7.5 μm. Results iOCT was feasible during vascular microsurgical procedures. In all scanned arteries a clear delineation of the physiological three layered vessel wall composition could be achieved. Pathological arteriosclerotic alterations of cerebral artery walls could precisely be demonstrated. Major superficial cortical veins conversely presented a mono layered composition. First in vivo measurements of vascular mean diameters were possible. Cerebral artery walls showed a diameter of 296 μm, tunica externa 78 μm, media 134 μm and interna 84 μm. Conclusion For the first time the microstructural composition of cerebral blood vessels could be illustrated in vivo. Due to an outstanding spatial resolution a clear definition of physiological and pathological characteristics was possible. Therefore, microscope integrated optical coherence tomography holds promise for basic research in the field of cerebrovascular arteriosclerotic diseases and for intraoperative guidance during microvascular surgery.


Results
iOCT was feasible during vascular microsurgical procedures. In all scanned arteries a clear delineation of the physiological three layered vessel wall composition could be achieved. Pathological arteriosclerotic alterations of cerebral artery walls could precisely be demonstrated. Major superficial cortical veins conversely presented a mono layered composition. First in vivo measurements of vascular mean diameters were possible. Cerebral artery walls showed a diameter of 296 μm, tunica externa 78 μm, media 134 μm and interna 84 μm.

Conclusion
For the first time the microstructural composition of cerebral blood vessels could be illustrated in vivo. Due to an outstanding spatial resolution a clear definition of physiological and pathological characteristics was possible. Therefore, microscope integrated optical coherence tomography holds promise for basic research in the field of cerebrovascular arteriosclerotic diseases and for intraoperative guidance during microvascular surgery.

Introduction
Imaging of cerebral vessel wall characteristics is fundamental to understand pathophysiological processes of cerebrovascular diseases and for intraoperative guidance during micro neurosurgery. Histology-albeit its' superior and unexcelled resolution-excludes an in vivo approach and is further constrained by numerous process-related artifacts, such as specimen shrinkage, tissue damage and tissue separation [1]. Traditional intraoperative imaging modalities such as ultrasound lack the necessary spatial resolution to delineate vessel wall characteristics as well as the possibility of microscope integration [2][3][4].
OCT imaging depends on near infrared light. In comparison to visible light-it delivers less energy. It has less electronic excitation potential of molecules and no thermal capacities. It is routinely used in ophthalmology since 30 years. No adverse effects or harming potential at the neuronal cell layer of the retina or in general are described so far [5].
In biological tissue penetrating depth is approx. 4000 μm and therefore suitable for cerebral vessel wall imaging [6]. It shows an exceedingly high spatial resolution ranging from 1-15 μm and therefore allows for identification of structural constituents to the extent of histology [7]. In ex vivo experimental set ups the correlation of intravascular OCT and histological findings of vessel walls could already be demonstrated [8][9][10]. In this context, extravascular OCT even seemed to exceed intravascular OCT in terms of image quality [11][12][13]. Physically depending on light, microscope integration is fairly simple if optics of microscopes are optimized for light transmission in near infrared spectral range. It then allows for rapid, real-time, three dimensional scanning of tissue in the field of view of the surgeon [14].
Various publications describe endovascular OCT for accessing coronary arteries during cardiac catheterization. Due to size and limited flexibility of fiber optics the intracranial use of intravascular OCT remains highly restricted. In case of the anterior Circle of Willis the intra carotid artery siphon in particular prevents intracranial passage [8]. Consequently, reports of neuro-endovascular OCT remain limited and mostly subject animal or ex vivo models [8,15]. Only a few case reports address in vivo vessel wall imaging of the infra-tentorial cerebral vasculature [16,17].
Due to these appropriate technical properties as well as the feasibility during micro surgical procedures, we here investigate microscope integrated optical coherence tomography as the first suitable in vivo imaging modality of cerebral vessel wall morphology.

Optical coherence tomography
If preparation and dissection of cerebral blood vessels was necessary for the surgical procedure OCT-scans were performed. The surgical microscope HS Hi-R1000G (Haag-Streit Surgical GmbH, Wedel, Germany) with integrated OCT-camera (OptoMedical Technologies GmbH, Lübeck, Germany) was used. A two axis scanner assured three-dimensional real time OCT scanning of tissue in the field of view. Volume scans-consisting of 30 B scans-each with a lateral scan width of 5-37 mm, an optical window depth of 3.1-4.2 mm and an axial spatial resolution 7.5 μm could be obtained.

PLOS ONE
OCT scanning was screened live on microscope monitors and via a head up display in the field of view of the surgeon. OCT volume scans, corresponding light microscopic pictures and light microscopic videos were further stored for post procedural analysis. The system was CE (conformity with the European Union guidelines) certified for intraoperative documentation of tissue structures. For detailed description on the setup see [19].

Image acquisition
The OCT scan site was defined by a senior vascular neurosurgeon with experience in OCT scanning. Focusing on the proximal branches of the anterior Circle of Willis with limitation due to the neurosurgical approach. An orthograde scan-angle and highest microscope magnification was intended. A corresponding light microscopic picture, video and three-dimensional OCT volume scan were recorded and stored with a data set defining OCT characteristics at time of scan.

Anatomical analysis
Data was sorted according image quality and richness of details. The aspect ratio of OCT scans needed to be adjusted according to microscope magnification during time of scan. This process followed a fixed protocol in every subject. These data sets were screened by experienced surgeons for anatomical analysis.

Distance measurements
A singular B-scan was selected from a set of volume scans. Anatomical structures were measured parallel to the optic path and under maximum augmentation. The center of the probe was selected to reduce scattering effects. Due to the spatial resolution of 7.5 μm final measurements were rounded to tens. Measuring was not adjusted according the index of refraction. For image processing ImageJ2 and Fiji were used [21,22].

Adverse effects
In one 17 yo female patient with resection of an extended right temporal mesial Ganglioglioma (WHO Grade I) the below mentioned vasospasm of an M2 segment occurred. The segment was scanned with OCT after opening of the dura mater. After extended microsurgical dissection of the tumor an intraoperative subarachnoid hemorrhage occurred followed by the below described vasospasm. Other modifications of vascular tone in temporal connection with OCT scans did not occur. None of the participants demonstrated global adverse effects. No prolonged wake up phase, novel neuronal deficits, headaches or epileptic seizures were clustered.

Cerebral arteries
In all scanned arteries the physiological three-layer composition of the vessel wall could be delineated. Pathological alterations like arteriosclerotic plaques, calcifications and circular inclusions could be further demonstrated. Light microscopy, corresponding OCT scan and illustration of these physiological and pathological characteristics as displayed by OCT are presented in Fig 1. The variety of cerebral vessel wall modifications in 10 exemplary OCT B-scans of 10 different cerebral arteries and patients are shown in Fig 2, https://osf.io/a3gx8/.

Cerebral veins
Scanned sylvian veins (SV) exhibited a thinner vessel wall (mean 53±23 μm, range 30-80 μm) than cerebral arteries. A stratum-like composition was not recognized. Arteriosclerotic modifications were not displayed. Regarding the vessel lumen, veins presented an intraluminal scattering of light. This phenomenon was absent in cerebral arteries without vasospastic conditions. An exemplary light microscopy, corresponding OCT scan and illustration of physiological, anatomical characteristics as displayed by OCT are presented in Fig 3, https://osf.io/a3gx8/.

Cerebral vasospasm
During preparation an incidental cerebral vasospasm could be scanned. Later analysis demonstrated an intra stenotic constriction of the vessel wall diameter, an enlarged diameter of the tunica media-presumably due to constriction of soft muscular tissue-and a mainly pre-but also post stenotic intra luminal scattering of light. Which we interpreted as an increased optical density due to a turbular flow in the periphery in relation to a laminar flow in the center of the lumen. A bird's eye view of a 3-dimensional OCT volume scan and corresponding OCT B-scans are illustrated in

Discussion
Various attempts using high-field and ultra-high-field MRI addressed non-invasive vessel wall imaging of intracranial cerebral arteries in vivo [23,24]. So far, spatial resolution remains limited with voxel sizes of 500-800 μm inhibiting a detailed depiction of vessel wall composition https://doi.org/10.1371/journal.pone.0276307.g001 [25]. As an intraoperative real-time imaging method ultra sound states the gold standard and is integrated in daily clinical practice. But spatial resolution capabilities also inhibits depiction of the vessel wall composition [2][3][4]. In contrast OCT shows an exceptional high spatial resolution-in our set up 7.5 μm-and penetration depth ranges from 1000-3000 μm [15]. In line with this resolution, experimental OCT analysis of vessel walls could priorly demonstrate coherence with histological findings [8][9][10]. No adverse effects are expected since scanning depends on near infrared light which has less irritation potential then the light spectrum of the operating microscope [26]. Due to its physical properties microscope integration is fairly simple [14]. These properties seem to be suitable to visualize and analyze the microstructural composition of vessel walls and integration of the technique in the microsurgical work flow.
This study presents for the first time detailed human vessel wall visualization of supra-tentorial cerebral vessels in vivo and real time. Due to the size and cross sectional design validity remains limited but no adverse effects were documented.

Cerebral arteries
OCT scanning revealed a three layered composition in all scanned arteries. Prior studies in coronary OCT verified these optical layers to match histological vessel wall strata [27]. Spatial resolution of OCT scanning met histology at intermediate magnification and imaging was well comprehensible. Pathological changes like arteriosclerotic plaques, calcification or intima hyperplasia could further be visualized and matched to the particular strata. Image quality seemed to outperform those of endovascular approaches [15,16]. In coherence to the literature we postulate that an increased optical density of blood in relation to air lead to increased light absorption during endovascular scanning decreasing image quality [15].

Vessel wall diameters
To the best of our knowledge, we here present the first in vivo measurements of cerebral vessel wall and single layer diameters. Absolute as well as proportional measurements varied widely.  Physiological, pathological and methodological factors might influence these measurements. Physiologically we are looking at individual variants of vascular size as well as individual vessel wall layer diameters. Furthermore, we are scanning different vascular tones. Vascular tone not only influences the absolute diameter of vessels but also the diameter of its three tunicae and the proportions among them. This physiological mechanism was remarkably illustrated by scanning of an incidental intraoperative vasospasm, see Pathological factors like arteriosclerosis lead to an increased diameter of tunica intima and in late stages to a diminished diameter of tunica media due to muscular atrophy influencing the vessel wall and vessel wall layer diameters [28,29]. This variability of pathological alterations is well demonstrated in Fig 2. Methodological factors could be slight deviation from an orthograde scan angel due to the limited surgical approach and-of course-the limited case number.

Cerebral veins
OCT scans of cerebral veins did not demonstrate a three-layer composition. The minor role of tunica media in veins regarding thickness, structure and function could be accountable. In arteries, this particular layer predominantly allowed to distinguish certain vessel wall layers. Rather thin walls in cerebral veins could further account for these findings. OCT scanning revealed venous wall diameters of approx. 30 μm. Which approaches our methodological limits.
Intraluminal light scattering was only found in veins and not in arteries. The lower blood velocity within veins could promote peripheral turbulent flow, according to Hagen-Poisseuille equation. Turbulent flow could lead to increased light absorption in relation to a central laminar flow. This would be in coherence with assumptions based on fluid dynamics and our observations in the vasospastic M2 artery segment [30,31]. Here, scattering of light was most prominent in the pre-stenotic segment, not visible in the intra-stenotic segment and presented again in the post-stenotic segment, see

Future perspectives
These results demonstrate that OCT can be used to study physiological and pathological alterations of vessel walls in vivo. One clinical application would be the field of vasa vasorum. These vessels are crucial in the pathogenesis of arteriosclerosis among multiple other cerebrovascular diseases [32]. Technically, OCT systems with further improved spatial resolution yet exist which could be able to delineate these vessels and adjacent vascular segments from an extravascular approach [33].
Microscope-integrated extravascular OCT should be further valued as an intraoperative imaging tool in the context of plastic, cardiac and neurosurgery. The highly demanding field of vascular neurosurgery with pathologies ranging from cerebral aneurysms, cerebral as well as spinal arterio-venous-malformations and Moyamoya disease could be targeted among others. In the latter, OCT could explicitly delineate the optimal site for vascular harvest and extraintra cranial bypass [34].

Conclusion
We here present the first in vivo imaging and measurements of cerebral vessel walls with extravascular microscope integrated OCT. Clinical adverse effects were not noted. Other optical based instruments of the neurosurgical operation room like the navigation system, infrared angiography or electro physiological measurements seemed not to be influenced.
It allowed detailed depiction of single arterial vessel wall layers, approaching spatial resolution of histopathology. Physiological changes of vascular tone with consecutive modification of the vascular lumen as well as diameters of singular arterial layers could be demonstrated clearly and in real time. Pathological conditions like arteriosclerotic modifications with intima hyperplasia, calcifications and local changes of vessel wall diameters could be depicted. Future studies should value microscope integrated OCT for basic research and as an intraoperative guidance tool in vascular surgery.